INTERPRETATION OF RESULTS
4.2 Data Collecting
Correlation Matrix
Factor analysis is a technique to identify and divide factors or variables which have the same characteristics into several variable groups. Factor analysis also conducts reduction or deleting to decrease number of variables.
After deleting the invalid statements based on comparison between r computation and r table, the researcher have to do reduction of variable. The table below is the fixed questionnaire after the researcher did reliability and validity testing
Table 4.4 Fixed Questionnaire
CORE BENEFIT
P1 Smartfren is the right choice as my internet service provider P2 Smartfren has answered my needs for internet service provider P3 I feel satisfy with Smartfren as my internet service provider P4 Smartfren offer excellent internet service and it is very helpful for
my daily activity
P5 Smartfren offer many benefits than the other internet service provider
FEATURES
P6 Features in Smarfren are easy to be used / user friendly P7 Features in Smarfren are enough to fulfill my needs
P8 Features in Smarttfren provide me with unlimited access to the internet
P9 Features in Smartfren is better than the other internet service provider
BRAND NAME
P10 I don‟t easily confused to compare brand name between Smartfren and another competitors
P11 Smartfren brand name represent characteristic of the product.
P12 Smartfren is an unique name
43
PRODUCT QUALITY
P13 Smartfren has succeed to prove their campaign „I Hate Slow‟
P14 Smartfren can deliver the service without any trouble or disturbance P15 Smartfren can give high speed internet service
P16 I never feel regret of using Smartfren as my internet service provider
P17 I choose Smartfren because of its‟ quality AFTER SALES SUPPORT
P18 I don‟t need to spend a lot of time in customer service if I have problem with Smartfren
P19 I feel comfortable with Smartfren customer service
P20 Smartfren use many kind of channel such as text message, email, internet, etc to keep me updated with Smartfren news.
P21 Smartfren customer services are highly trained to deal with customer‟s problem
P22 Smartfren always upgrade and renew services that delight customers.
WARRANTY
P23 Smartfren mention the warranty points clearly P24 I can easily claim warranty in all Smartfren store P25 I feel protected by warranty of the product
P26 I can understand each point of the warranty easily
Reduction of variable can be done in several steps. The researcher decide which factors are appropriate to be used in next analysis is by using KMO and Bartlett‟s Test.
Table 4.5
KMO and Bartlett’s Test
Kaiser-Meyer-Olkin Measure of Sampling
Adequacy. .780
Bartlett's Test of Sphericity Approx. Chi-Square 981.829
Df 325
Sig. .000
Source: SPSS 16 and Primary Data
44
From the table above, the value of KMO MSA (Kaiser – Meyer – Olkin Measure of Sampling Adequacy) is greater than 0.5. Based on theory if value of KMO MSA is greater than 0.5, the researcher can continue the analysis. It also means that sampling technique in the study can be applied.
The next step is evaluating Anti-Image Matrices to decide which factors are appropriate to be included in the next analysis. There are 6 variables in this research, which are core benefit, feature, brand name, product quality, after sales support, and warranty. Each variable consist of five questions. With details as follow:
Statement 1 – 5 : Core Benefit of Product Statement 6 – 9 : Product‟s Features Statement 10-12 : Brand Name Statement 13-17 : Product Quality Statement 18-22 : After Sales Support Statement 23-26 : Warranty
The table of Anti Image Matrices is shown below:
Table 4.4 Anti-Image Matrices
No Statement MSA Value
P1 Smartfren is the right choice as my internet service
provider 0.858
P2 Smartfren has answered my needs for internet service
provider 0.810
P3 I feel satisfy with Smartfren as my internet service
provider 0.855
P4 Smartfren offer excellent internet service and it is very
helpful for my daily activity 0.859
P5 Smartfren offer many benefits than the other internet
service provider 0.879
P6 Features in Smarfren are easy to be used / user friendly 0.755 P7 Features in Smarfren are enough to fulfill my needs 0.795
45
P8 Features in Smarttfren provide me with unlimited access
to the internet 0.589
P9 Features in Smartfren is better than the other internet
service provider 0.831
P10 I don‟t easily confused to compare brand name between
Smartfren and another competitors 0.647
P11 Smartfren brand name represent characteristic of the
product. 0.630
P12 Smartfren is an unique name 0.614
P13 Smartfren has succeed to prove their campaign „I Hate
Slow‟ 0.861
P14 Smartfren can deliver the service without any trouble or
disturbance 0.728
P15 Smartfren can give high speed internet service 0.742 P16 I never feel regret of using Smartfren as my internet
service provider 0.824
P17 I choose Smartfren because of its‟ quality 0.890 P18 I don‟t need to spend a lot of time in customer service if I
have problem with Smartfren 0.667
P19 I feel comfortable with Smartfren customer service 0.775 P20
Smartfren use many kind of channel such as text message, email, internet, etc to keep me updated with Smartfren news.
0.810 P21 Smartfren customer services are highly trained to deal
with customer‟s problem 0.796
P22 Smartfren always upgrade and renew services that delight
customers. 0.734
P23 Smartfren mention the warranty points clearly 0.692 P24 I can easily claim warranty in all Smartfren store 0.823 P25 I feel protected by warranty of the product 0.787 P26 I can understand each point of the warranty easily 0.757 Source: SPSS 16 and Primary Data
Mueller et al (1978) states that if MSA (Measure of Sampling Adequacy) Value is less than 0.5, the variable must be deleted and it is not an appropriate factor to be entered in the next step of analysis.
From the table above, all the statements have MSA value greater than 0.5. It means, all the statements can be entered to the next step which is factor analysis.
46
Factor Extraction
The extraction of manifest variable is very important to figure out the latent variable.
Principal Component Analysis is used to generate the last factor extraction statistic.
By relying final statistic, there are three components can be identified which are communality, eigen value, and cumulative percent of extracted factors.
Eigen value is use to determining how many latent variables that will be generated, which is means that if the eigen value is more than 1, so it is considered as significant. The community of variables shows the variance proportion of the variable, which can be explained in generated factors. The range of communality of variable is between 0 to 1. The bigger value, the better it is because the variable become easier to explained by generate factor.
Based on Total Variance Explained table, there are 4 values that are shown as follow:
Communality shows the variance proportion of variable toward the whole factors.
Eigen value must more than 1, it shows the total variance on each factors. The first factor has the biggest eigen value which is 8.312. Based on the calculation, there are 6 factors that have eigen value exceed more than 1 percentage of variance, which show us the component number 1 is the highest percent variance, with the value is 31.969%.
The total variance is strong, which is 67.109%. It means that 32.891% of variables cannot be presented or become the error of this study. Some factors that is difficult to be interpreted because there are too many manifests that have exceeding values in more than one factor. Therefore, rotated component matrix is needed to go on the next process.
47
Table 4.7
Component, Eigen Value, %Variance, Cumulative %
Component Eigen Value % Variance Cumulative %
1 8.312 31.969 31.969
2 3.163 12.165 44.134
3 2.076 7.983 52.117
4 1.576 6.062 58.178
5 1.225 4.710 62.889
6 1.097 4.220 67.109
Source: SPSS 16.0 and Primary Data
Rotated Component Matrix
Rotated component matrix is use to get simpler factor structure which will make the variables interpretations become easier. From extraction result interpretation of matrix start from the left side (factor 1) to the right side (factor 6)
In the study orthogonal Varimax is used in rotated component methodology.
Orthogonal Varimax is used to rotate the beginning factor from extraction result, so at the end, it will create the rotated result where one column closer to zero. The rotated component matrix can be seen below:
Table 4.8
Manifest Variable and Factor Value
Factor Manifest Variable Factor Value
1 P3
P2 P1 P5 P4
0.848 0.817 0.814 0.673 0.612
48
2 P14
P13 P15
0.797 0.747 0.746
3 P22
P20 P8
0.707 0.689 0.650
4 P25
P26 P23 P24
0.838 0.702 0.673 0.604
5 P11
P12 P10
0.858 0.756 0.709
6 P18
P19
0.677 0.647 Source: SPSS 16.0 and Primary Data